An intelligent decision support system for part launching in a flexible manufacturing system

The effective control of flexible manufacturing systems (FMS) poses a major challenge to designers and the research community. In this paper, the fram ework of a unique decision support system to assist in the control of a FMS through intelligent part-launching is introduced. The proposed system make s use of a new heuristic method based on the pull concept, and a neural net work trained with data from a system simulation model, which in turn provid es a decision about choosing a part to be launched into the FMS. The perfor mance of the proposed heuristic method was tested against another popular h euristic method normally used in existing FMS via discrete event simulation models. Data sets from the simulation program of the proposed system were used to train the neural network, which can then be used on the shop floor for realtime decision support. The output from the trained neural network i s based on the same principle as the proposed heuristic method used in the simulation program. Therefore, the decision support provided by the neural network can be used for improved shop floor control.